Yodeller

One of the great things working on a consulting company is that there are always new services released. Even though they are projects for our customers it's still thrilling to finally get your work out there for the wider audience.

It was bice back when I was still working on those projects, but it has been years since I have last time done that. Sure I get to participate on those launches and celebrate all the success, but it's not the same when you haven't been directly involved in those projects.

On my recent roles my "releases" have mostly been new processes and policies. Not that thrilling. But today I got a feel of that omd glory when I was able to announce the release of our internal AI platform in our company meeting.

I've been working on that project mostly alone and while the userbase won't be huge I believe they will at least be enthusiastic about it. At least the reception and first impressions from our employees were positive.

Until now it has been my project, but bow it's time to hand it over to others to continue to work on it. I just show them the way. Of course I will still retain the ownership of the project and have a say on the direction, but I doubt I will have that much time to put into the further development of it.

A good cliffhanger can keep me reading for hours. Some authors have really mastered the skill of keeping their readers captivated. But some take it a bit too far, weaving too many story arcs that intervene each others.

Building up the suspense towards the end of the chapter keep you continuing to the next one just to find out what will happen. But if the next chapter isn't going to resolve that suspense it gets a bit tiring. It might answer the open questions from the few chapters back, but then you need to let go of the desire to know what happens next on the story of the chapter you just finished. You also need to recall what happened in that previous chapter the next one will continue on.

This is still bearable if there are only a few different story arcs going on simultaneously. But I have read books where there are half a dozen or even more things happening "at the same time". It just gets frustrating having to read a hundred pages before you get your answers. By that time you have a whole bunch of new ones waiting to be unfurled.

Privacy is big thing these days when it feels like every other service is just trying to harvest your data instead of just serving you. Their product is just a tool to get you hand over that data and the real product is you.

This has led to some people being extremely conscious of their privacy, even with the services that don't necessary collect their data. Or at least intend to sell it. Of course there is also the "traditional" privacy of being concerned who can access your data in the first place. Especially for the more personal and sensitive things.

In our company we try to respect the privacy of our employees. But this often leads to conflicts especially with the security related matters. People don't like to be monitored, but ensuring security there needs to be certain level of monitoring still in place.

Company provided tools and services can't be assumed to be private in the same level as privately used ones. After all they are company provided and intended for people to be able to do their jobs. We don't have strict policies against using company provided tools for private things as long as certain caution is being taken care of. But that freedom comes with the price of privacy.

While deploying our internal A.I. service I have been looking for GPU options to speed up the local models I want to enable as an option along the premium services. There are definitely no shortage of options, but they all come with quite hefty price tag.

There are turnkey solutions where tou can just select a model and deploy it to the cloud without worrying the underlying infrastructure. The ease sure comes with a price. These were the first options I ruled out. There are also middle ground options where you still need to take care of some of the infrastructure that are somewhat cheaper.

Going with the cheapest option does require quite a lot of work getting the setup right, but the benefit will be the savings from the running costs. It also enables running more models when you can actually run those on the same instance instead of having an individual instance for each model.

Setting such custom solution takes it's own time. So to not to delay the release of the service the local models feature have to be postponed, or launces with much slower CPU interference.

As foretold the new Dune prequel series' first episode was just released. Even though the episodes will be released one at the time a week apart I couldn't resist watching the first episode right away.

It's too early to judge the series based on just one episode. The first one in a series is always a bit slow start. Setting the background, introducing the characters and setting the foundations of the main plot(s) takes time.

The series is loosely based on (one of) the prequel trilogies written by Frank Herberts son Brian. After reading the two sequels written by him to the original series I haven't read any of those prequels so I can't really say how well this series complies with the source materials.

The one thing I did notice is that they didn't once reference the war against the sentient machines by it's name. Maybe it's one of those copyright things when the producers don't own rights to all the materials, or maybe it's just to save the viewers from some additional obscure names which there are already enough now. It could also be due to the current trend of trying to avoid certain sensitive words, like the jihad in the Butlerian Jihad which apparently has ended up in the banned words list.